|23rd February 2021, 10:30 PM||#72|
Join Date: Mar 2017
Bob Codell's book is helpful in Ch. 24 (1st Ed) or Ch. 28 (2nd Ed) "The Negative Feedback Controversy".
Is global feedback inherently bad? I'd say the short answer is No.
It isn't Spectral Growth Distortion or re-entrant distortion from Baxandall's analysis that global feedback inherently bad. Bob Cordell's book shows it is irrelevant to our audio amplifiers; transistors invariably use degeneration and this pushes the level of local feedback over the Spectral Growth bump (1st Ed. Fig. 24.4, 2nd Ed. Fig. 28.4). So adding global feedback around the usual follower output stage doesn't cause a Spectral Growth bump (see Bobs 1st Ed. 24.5, 2nd Ed. Fig. 28.5). He shows any amount of global feedback can be used, starting from 0dB in our power amps, without upsetting the harmonic structure of a typical follower power stages.
And it isn't global feedback that make an amplifier oscillate - even single stage amplifier can be made to oscillate - like the Colpitts oscillator which is a single stage emitter follower that is sufficiently capacitively loaded.
There are a few other debunked reasons for not using global feedback in Bob's book that I won't mention here.
The main problems I believe with global feedback are
1) recovery from clipping, and
2) sharpness of the clip region.
Bob Cordell says amplifiers sound different because they recover from clipping differently. Bob also says soft clipping power amplifiers can be over-driven more than hard clip amps before the clipping becomes obvious and this gives soft-clip amps a better listening experience (my paraphrase of Bob).
Bob built a clip-power meter and demonstrated it at a HES workshop and showed under typical listening levels power amps will clip even though most are unaware it of it CordellAudio.com - Home Entertainment Show 2007. Many power amp designers and builders assume clipping doesn't happen on their system and therefore think soft-clipping isn't a big issue. Actually, it depends a lot on your speaker's sensitivity; if your amp is in the usual 100W-200W range you should consider adding soft-clip to your amp if your speakers have less than 90dB/W/m sensitivity.
Bob says not using any global feedback won't necessarily give an amplifier soft-clipping, nor will it guarantee its free from clip-recovery problems. So using global feedback is not the culprit per se for killing soft-clipping or the clip-recovery problems. Global feedback just makes them worse. But there's no inherent problem here with global feedback.
When you look at the equations for feedback in a power amplifier there's no distinction in the equations when global feedback is used over local feedback - the equations don't differentiate between the two types of feedback. Obviously when more stages are enclosed by global feedback the number of poles and zeros that are enclosed by this feedback increases, which means the difficulty of maintaining an adequate gain and phase margin increases. But opting for no global feedback does not make things much easier - eg read Arto Kolinummi's thesis "Audio Power Amplifiers - towards inherently linear amplifiers" here on designing amplifiers with no-global feedback.
Last edited by IanHegglun; 23rd February 2021 at 10:45 PM.
|Thread||Thread Starter||Forum||Replies||Last Post|
|XLR to RCA adapter causes gross amounts of distortion...||shredhead||Analog Line Level||13||7th June 2015 02:00 PM|
|Class D amp distortion causes||zigzagflux||Class D||20||13th November 2014 12:10 PM|
|Amp causes speaker distortion||blubberboy||Solid State||4||11th August 2013 02:49 PM|
|Determining causes of distortion||shzmm||Solid State||4||9th February 2010 01:58 AM|
|What causes distortion?||matthias||Parts||1||5th June 2004 04:47 AM|
|New To Site?||Need Help?|